This project focuses on two recent hypotheses of early-emerging social cognitive mechanisms shown to be impaired in individuals with autism. The first hypothesis states that when viewing naturalistic social situations, individuals with autism exhibit aberrant patterns of visual fixation and that these patterns predict the participants' relative levels of social competence. This hypothesis originated from our research using eye-tracking technology to measure fixation times on different aspects of social scenes. The second hypothesis states that individuals with autism have a decreased ability to impose social meaning on ambiguous visual stimuli. This hypothesis originated from our research in which participants with autism had difficulty making social attributions to geometric animations depicting typical social interactions. We request 5 years of support to examine these hypotheses in 60 infants suspected of having autism spectrum disorders (ASD) seen longitudinally at age 24 to 36 months (Time 1) and again at age 48 to 60 months (Time 2). This group will be matched on nonverbal MA to 30 children with non-autistic developmental delays (DD) also followed longitudinally. ASD and DD groups will be matched to typical controls cross-sectionally (N1=40 in Time 1 and N2=40 in Time 2). In Specific Aim 1 we plan to use eye-tracking technology to study visual scanning patterns during viewing of naturalistic social interactions presented as scenes of children at play and as facial close-ups of an engaging adult. In Specific Aim 2 we plan to use eye-tracking technology to study the capacity for imposing social meaning upon ambiguous representations of people (people rendered as moving point-light displays). For both Specific Aims, data will be analyzed cross-sectionally at Times 1 and 2. At both times we plan to study the relationship of visual fixation patterns to diagnostic group membership and to standardized outcome measures of social and communicative competence. In addition, the longitudinal component of the study creates an opportunity to test the predictive power of aberrant visual fixation patterns at Time 1 relative to confirmatory diagnostic group membership and outcome measures at Time 2. Apart from providing a unique window into the ways in which young children with autism search for meaning when confronted with social situations, our longer-term goal is to develop our eye-tracking paradigms into performance-based measures capable of (1) identifying children with autism in the first two years of life (an important goal given the proven benefits of early intervention), and (2) defining a continuum of social competence (a critical component in genetic research of varying manifestations of autism). This project builds on funded work using similar paradigms both with older individuals with autism and with an animal model of autism (non-human primates with mesiofrontral-limbic ablations).